Evaluation 111

In this section the evaluation of the evaluation criteria described in chapter 2.4.3 are performed, to secure the quality of the study. The evaluation of the quality criteria are followed by a method evaluation that aims to evaluate the chosen research method.

8.2.1 Evaluation criteria Credibility

During the research process we have worked towards creating a high credibility. This has been done in multiple ways. First of all we have used different way of triangulation to increase the credibility of our findings. This triangulation process has consisted of several steps, where the first step was to include different cases in the case study, as well as studying the cases from multiple points of view. By doing so the credibility of our findings increased as it made it possible to use data from different cases and views to create a result that was in line with all findings.

The above mentioned triangulation in combination with an analysis of the triangulated material resulted in a preliminary result. To further increase the credibility of the preliminary result, another type of triangulation was used. This triangulation process was performed by comparing the preliminary results against theory, i.e. a theoretical validation, as well as through the use of a semi-structured interview with a person involved in diffusion projects in developing countries, i.e. an empirical validation. The aim with both the theoretical and the empirical validation was to increase the credibility by comparing the result with previous theoretical findings as well as against a practitioner’s view.

Even though we have been working hard towards increasing the credibility of our result, the study has been performed in Thailand, where the general English knowledge is low. This might have affected the credibility of our findings in a negative way, as it forced us to use an interpreter when performing our interviews. All our questions and every answer had to go through our interpreter and as such we had to rely on that all questions and answers had been properly understood. To offset this difficulty in some ways we used a highly educated interpreter that had a master’s degree, with which we spent an entire day discussing our research and goals to make sure that we had a unified view of what we were trying to achieve.

Transferability

As we have conducted a qualitative case study it can be difficult to ensure a high transferability of our findings. However as the aim of our result was to be applicable in developing countries in general and not only in Thailand, it was important for us to work against creating as high transferability as possible, given our circumstances.

One of the methods used when trying to create a higher transferability was to give a detailed description of our collected material and cases. This was in line with Oates (2006) who describes that a ‘thick’ description can help to create a higher transferability because it enables the readers to judge whether their own situation of interest has similar features so that the findings could be relevant there too. The fact that we have conducted a case study with multiple cases has also helped to strengthen the transferability as the findings are not as narrow as if the study only had been based on a single case.

Internal reliability

As we are two researchers we have been working towards creating a high internal reliability in our results. To increase the internal reliability of our findings, we have continuously throughout the entire research process worked together and always had discussions about our research to make sure that we have had a shared view about our studied area and findings. In the case of different opinions we have had discussions about the differences to create a unified view, something that was important to secure the quality of our findings and results.

One limitation connected with internal reliability and our study is that the entire study has been conducted with the help of an interpreter. Even though we have had discussions about the aim of our views with the interpreter, one cannot be totally sure about that the interpreter shared the same view and standpoint as we did.

Confirmability

As the study has been a qualitative case study, it can be difficult to ensure complete objectivity. However, throughout the entire research process we have tried to leave our personal values and opinions out of the research as much as possible. This has been an important aspect as we have created guidelines that can be applied to more areas than the studied case.

When analyzing and drawing conclusions about the collected material, we have tried to leave personal opinions and values out of the work process to make sure that the these opinions and values did not affect the collected material. As we have performed an inductive research, it has also been important for us to make sure that theories in our theoretical framework has been left out and not affected our results. This was important as it enabled the preliminary guidelines to be based solely on our collected empirical material, and as such be as representative as possible for our studied cases.

As with the evaluation criterion internal reliability the use of an interpreter can affect the criterion of confirmability, because we have had no control over how the interpreter asks the questions as we have no knowledge about the indigenous language. It has also been difficult to not be affected in any way by the theories included in the theoretical framework.

Generalizability

During the research our focus has been on creating general guidelines that are applicable on developing countries in general, not only in Thailand. It has therefore been important for us to work towards creating as high generalizability as possible, given our circumstances and type of research strategy.

To ensure that the developed guidelines can be used in different settings than the studied one, different techniques have been used to increase the generalizability of our finding. First of all we chose to include multiple cases in our case study, which added some breadth to the findings as it made it possible to compare the findings from the different cases, and create guidelines that were based on more than one data source. To add another dimension we also chose to perform interviews with people from different positions as well as interviewing both sides of a diffusion project, i.e. donors and recipients. By doing so it was possible to get the views from different sides of a project, which made it possible for us to identify potential gaps and similarities between the two sides. When creating the guidelines we tried to make them as general as possible. This was done by indentifying the underlying problems to what had been said, something that made the guidelines more general rather than specific for the studied cases. One can also argue that the fact that we have conducted nine interviews increases the possibility of generalizability as it led to large amounts of data about the studied cases.

Even though we have worked hard to create a high generalizability, it is important to understand that the fact that we have conducted a case study, which is narrow and focuses on depth rather than breadth, could have had a negative effect on the generalizability. This as the case studied may be hard to transfer to other cases. One must also consider that Thailand in many ways are more well developed than many of the other developing countries, and can longer be considered as a typical developing country (for details see chapter 2.4.1.1). The fact that the situation in Thailand might differ slightly compared to other developing countries could have had a negative impact on the guidelines since challenges and problems unknown to Thailand might exist in other countries. This problem was hopefully alleviated some by the fact that we conducted both a theoretical and empirical validation of the results and findings, and complemented the guidelines if any contradictions existed.

8.2.2 Method evaluation

We have conducted a qualitative study, which has suited the purpose of our study well as we had to collect rich and detailed data to gain a deeper understanding about the studied cases. This type of data was required for us to be able to answer the research questions of the study. This would not have been possible if we instead had taken on a quantitative approach as we would not have been able to gain the knowledge and understanding about the situation at schools in the rural areas of Thailand.

Throughout the entire research process we have been using an interpretivistic perspective which has led to us continuously trying to gain an understanding about the underlying reasons behind what was being said during the interviews. To gain this understanding we had to interpret what was being said at all times which might have created problems as misinterpretations may have occurred, and this is especially evident since the study was

performed in Thailand with the use of an interpreter. However, we tried to minimize the interpretation errors by conducting both a theoretical and an empirical validation of all findings and results.

The study has been conducted in an inductive manner, which led to us creating the guidelines based solely on the empirical material collected from the interviews. This fitted the purpose well as few studies about diffusion of e-learning to elementary schools in rural areas of developing countries have been performed and it was therefore important for us to create our own theory. This fact also made it evident that it would have been difficult to conduct the study with a deductive approach due to the lack of existing theories to base the study on. As a new theory was created from the collected empirical material it was important for us to test and validate our findings and result. The test was performed both empirically and theoretically, which made it possible to validate our findings from a broader perspective. As the strategy of the study have been to first collect material in an inductive manner, and based on that create a new theory regarding the diffusion of e-learning in developing countries, it was important for us to start by validating our findings against theory. By validating the findings against theory it was possible to further strengthen them as a preparation for a final empirical test. Further empirical tests would have been needed, but as that was outside the scope of the study, only two validation tests have been performed on the findings. These tests were performed to secure the quality and authenticity of the findings and results.

The theoretical framework was included and used in the study for two purposes; as a base when creating the interview guides and to validate the findings through the theoretical validation. The theoretical framework was used when creating the interview guides for guidance about which aspects and areas that had to be considered. By doing so it was possible to secure that no important aspects would be left out during the interviews. The fact that the theoretical framework also was the basis for the theoretical validation made it important that the chosen theories fitted the purpose of the study and the research questions well. This is something that we in retrospect feel that we succeeded with as a lot of the collected empirical material and the analysis of that empirical material seems to be in line with the theoretical framework.

During the study we have been using semi-structured interviews to collect all empirical material. This has suited the purpose of the study well, as it made it possible for us to get some structure on our interviews to make sure that we did not miss any important aspect, as well as giving us some flexibility to add questions or change questions to fit the respondents. It is even more evident that semi-structured interviews were a suitable choice as the study was performed in Thailand with an interpreter. If we instead had used e.g. unstructured interviews or focus groups it would have been difficult to ensure that the collected material was of high enough quality. By using unstructured interviews we would have had to rely solely on that our interpreter would be able to interpret and reproduce everything that had been said during the interview in order for us to be able to collect material of high quality. The same thing can be said about focus groups, but adding the fact that the interpreter would have had to be able to act as the moderator of the focus groups, which would have made us loose to much control over the situation and as such the ability to secure the quality of the material.